We conducted a population-based cohort study of the receipt of prescriptions for inhaled corticosteroids, before and after introduction of the income-based drug policy, among school-aged children (5 to 15 years old) treated for asthma. The data were obtained from 4 computerized databases maintained by the Manitoba Health Service Insurance Plan (MHSIP): registration files, physician reimbursement claims, hospital discharge abstracts and prescriptions dispensed in retail pharmacies. The MHSIP registration file contains a unique numeric identifier for every registrant, which allows linkage of the health care records. MHSIP's prescription and health care administrative databases have been shown to have good reliability and validity.13,14
The study protocol was approved by the Human Research Ethics Committee, University of Manitoba, and permission to access the data was obtained from the Manitoba Health Access and Confidentiality Committee.
Children were selected for inclusion if they had received prescriptions for asthma drugs in the 1-year period before and the 2-year period after introduction of the income-based policy on Apr. 1, 1996, as follows: at least 1 prescription for a bronchodilator, inhaled corticosteroid or cromone, ketotifen, or oral corticosteroid, in conjunction with at least one physician visit or admission to hospital related to a diagnosis of asthma or bronchitis; or, in the absence of such a diagnosis, at least 1 prescription for an inhaled corticosteroid or cromone or for ketotifen in conjunction with a bronchodilator, or at least 2 prescriptions for a bronchodilator. Because our definition excluded children with one-time bronchodilator prescriptions and no diagnosis of asthma,15
it was more stringent than definitions used by others.16
In the absence of direct measures of household income, we used 1996 census information on average household income reported for enumeration areas to rank households into quintiles, from the 20% of the population in the lowest-income neighbourhoods to the 20% of the population in the highest-income neighbourhoods.17
Three groups were identified: children in households receiving prescriptions reimbursed in full by the income assistance and treaty First Nations prescription programs, as defined in the prescription database (the non-Pharmacare group), children in households receiving Pharmacare benefits that were located in neighbourhoods in the lowest-income quintile (the low-income Pharmacare group) and children in households receiving Pharmacare benefits that were located in neighbourhoods in the 4 higher-income quintiles (the higher-income Pharmacare group). The non-Pharmacare group was the comparison group, because there was no change in the drug reimbursement policy for this group.
To diminish confounding by disease severity,18
the children were stratified by asthma severity, as derived from the asthma prescription drug profile19
and the history of hospital admissions, as follows: mild to moderate asthma was defined as use of bronchodilators with or without inhaled corticosteroids or cromones, and severe asthma was defined as high use of bronchodilators (greater than the 90th percentile of doses) with use of inhaled corticosteroids or cromones or admission to hospital for asthma, or use of oral corticosteroids. Children who did not receive any inhaled corticosteroids were classified as having severe asthma if they used oral corticosteroids or high doses of bronchodilators. The severity measure was found to have good reliability (kappa = 0.82) and validity through its association with other markers of severity, such as admission to an intensive care unit.20
By applying the severity criteria to health care data from before and after introduction of the income-based drug policy, we obtained 2 subgroups: 6612 children with mild to moderate asthma before and after introduction of the income-based policy (referred to here as stable, mild to moderate asthma) and 1420 children with severe asthma before and after introduction of the new policy (referred to here as stable, severe asthma). Children with decreasing (n
= 1223) or increasing (n
= 1448) severity of asthma over time were excluded because of the difficulty in distinguishing changes in prescriptions for inhaled corticosteroids that were secondary to changes in asthma severity from changes associated with the drug benefit policy.
Two measures of inhaled corticosteroid use, computed for the 1-year period before and the 2 consecutive 2-year periods after the policy change, were determined: the proportion of children who received a prescription and the mean number of doses per child-year among children whose prescriptions were filled, derived from the prescription quantity and the standard unit sizes of inhalers. Inhaled corticosteroid drugs included beclomethasone, budesonide, fluticasone, flunisolide and triamcinolone. Comparisons between time periods of the likelihood of receiving an inhaled corticosteroid prescription, adjusted for monthly variation and stratified by income status and asthma severity, were assessed with Mantel–Haenszel odds ratios (ORs) and 95% confidence intervals (CIs). The Breslow–Day test of heterogeneity was applied to assess whether the likelihoods for the non-Pharmacare and the low- income and higher-income Pharmacare groups were statistically different. A split-unit analysis, reported as least-square means and 95% CI, was conducted to determine the mean number of inhaled corticosteroid doses before and after the policy change in relation to income status. Assuming an α of 0.05, a β of 0.2, and one-sided and paired (before and after) comparisons, 496 children were required per stratum to find an OR of 0.80, and 371 children were required per stratum to detect a 10% decrease in dose.